Phenopy Versions Save

Overview

This version modernizes most of the codebase, including a refactor to ensure compatibility with Python 3.9

What's Changed

• Update to Python 3.9
• Includes a pyproject.toml file for use with poetry (removal of setup.py)
• Now relies on gensim = ^4.3.0
• Replaces flake8 with ruff
• Replacement of unittest with pytest
• Removal of CLI test workflow
• Replacement of scipy.stats.truncnorm function with custom truncated normal-simulation functions
• Addition of docstrings and type-hinting
• More flexible versioning for fire, networkx, numpy, pandas, scipy than v0.5.4

Known Issues

• phenopy v0.6.0 relies on Python3.9, which is expected to reach end-of-life on 2025-10-05.

Compatibility

Requires >= Python 3.9.

Contributors

• Stephen McGee (@mcgeestephen)

Release Date

Version 0.6.0 was released on June 16, 2023.

Overview

This version incorporates changes to handle new column headers in the phenotype.hpoa file released for the Human Phenotype Ontology (HPO), updates dependencies, and removes certain code elements and dependencies.

What's Changed

• Can now handle new column headers of phenotype.hpoa file from 4/6/2023 release of HPO
• Merge pull request #92 from GeneDx/develop by @mcgeestephen
• Bump requests from 2.25.1 to 2.31.0 by @mcgeestephen in https://github.com/GeneDx/phenopy/pull/92
• Removed likelihood of molecular diagnosis code by @mcgeestephen (likelihood.py, test_likelihood.py)
• Removed likelihood_moldx module from phenopy/__main__.py
• Removal of joblib and lightgbm dependencies by @mcgeestephen
• Versions explicitly stated in setup.py for fire, networkx, numpy, pandas, scipy

Known Issues

• phenopy v0.5.4 relies on Python3.7, which is expected to reach end-of-life on 2023-06-27.
• vocab was deprecated from KeyedVector and will throw an AttributeError for applications relying on newer gensim versions (>=4.0.0)

Compatibility

Requires >= Python 3.7.

Contributors

• Stephen Mcgee (@mcgeestephen)

Release Date

Version 0.5.4 was released on June 6, 2023.

Handles the new hpoa file format

Fixes installation by removing code from __init__.py and distributes data dependencies when phenopy is installed. Breaking changes occurred by removing code from __init__.py

Breaking Changes: If you use from phenopy import generate_annotated_hpo_network, it should now be from phenopy.build_hpo import generate_annotated_hpo_network.

Use the pipenv environment to build dist/* and upload with twine.

Major:

• Source code for the phenoseries experiment
• Word2vec scoring functionality
• Use custom disease records for annotation
• Added likelihood of molecular diagnosis functionality

Minor:

• Added Jaccard scoring method
• moved multiprocessing out of main so library usage takes advantage of multiple cpus
• Additional unit tests

Adds functionality to score two sets of HPO terms without creating records.

This fixes a bug when parsing the updated hpoa format. The new format simply adds a leading hash to the line containing the field names.

• update IC of root node to 0.0
• update to use new HPO disease disease annotation

*remove rounding *bibtex